An effective greedy randomized adaptive search procedure-based hybrid metaheuristic for the distributed no-wait flowshop scheduling problem
Journal of Computational and Applied Mathematics, cilt.488, 2026 (SCI-Expanded, Scopus)
- Yayın Türü: Makale / Tam Makale
- Cilt numarası: 488
- Basım Tarihi: 2026
- Doi Numarası: 10.1016/j.cam.2026.117834
- Dergi Adı: Journal of Computational and Applied Mathematics
- Derginin Tarandığı İndeksler: Science Citation Index Expanded (SCI-EXPANDED), Scopus, Applied Science & Technology Source, Compendex, INSPEC, MathSciNet, zbMATH, Academic Search Ultimate (EBSCO), Engineering Source (EBSCO)
- Anahtar Kelimeler: Combinatorial optimization, Distributed no-wait flowshop scheduling problem, Greedy randomized adaptive search procedure, Hybrid metaheuristic, Variable neighborhood descent
- Ondokuz Mayıs Üniversitesi Adresli: Evet
Özet
This study investigates the distributed no-wait flowshop scheduling problem (DNWFSP), a complex scheduling variant that arises in distributed manufacturing systems with multiple identical factories and strict no-wait constraints, where jobs must be processed continuously across machines without delay. The problem requires jointly deciding the assignment of jobs to factories and determining their no-wait sequences within each factory, making it computationally challenging and highly relevant to industries such as chemical, food, and pharmaceuticals. To address this problem, a hybrid metaheuristic combining the Greedy Randomized Adaptive Search Procedure (GRASP) and Variable Neighborhood Descent (VND) is developed and referred to as GRASP_VND, balancing diversification and intensification during the search process. It features a time-adaptive GRASP construction phase, an adaptive destruction–reconstruction mechanism for solution perturbation, and a structured multi-neighborhood local search (MN_LS) that explores intra-factory swaps, block relocations, and inter-factory job exchanges. Moreover, Bayesian Optimization is employed to fine-tune key decay parameters, enhancing the algorithm's adaptability across different instance characteristics. Computational results on 720 benchmark instances indicate that the proposed GRASP_VND algorithm performs competitively compared with several existing algorithms, including Iterated Local Search (ILS), Iterated Greedy variants (IG_RNS, IG_VND, and IG_VNS), Multi-Neighborhood Adaptive Tabu Search (MNATS), and the Novel Evolutionary Algorithm (NEA), in terms of solution quality.